Effects of speaker gender on speech recognition by hearing impaired listeners

Abstract

Audiologists and clients often claim that female speech is less intelligible than male speech for subjects with bilateral high‐frequency sensorineural hearing losses. To examine the extent to which source characteristics (voice pitch) and gender‐related resonance characteristics contribute to these speech recognition difficulties, two men and two women each produced three sublists of the modified rhyme test (MRT). Within gender groups, one speaker had a relatively low voice pitch, one a high pitch, such that the high‐pitched male voice was approximately the same as the low‐pitched female voice. Stimuli were presented to listeners at 30 dB SL in quiet, in speech noise (S/N = + 12 dB), and in speech babble (S/N = + 12 dB). Overall results showed significant differences in identification of consonants across speakers. However, contrary to expectations, performance was worst for the male with the low voice pitch, intermediate for the two female speakers, and best for the high‐pitched male. There appeared to be gender‐related differences in the type of consonant errors.